There are 191 pieces 0207/0204 resistors, Jacco 
and several hundred in protection and input boards...I like complicated circuity It is not my challenge to Krell, it is only example of configuration... This amp is still under developement...I will show you, when it will be ready, be patient...Only in short : it is mosfet amp with error correction, 400 W / four Ohms...But it will be not amp for begginers...
and several hundred in protection and input boards...I like complicated circuity It is not my challenge to Krell, it is only example of configuration... This amp is still under developement...I will show you, when it will be ready, be patient...Only in short : it is mosfet amp with error correction, 400 W / four Ohms...But it will be not amp for begginers... 
amplifierguru said:
Well, so much for the lack of finger pointing and the gentle feel of this thread.
Western Nations? I see you don't even acknowledge what country you reside in.
I actually like the way my Krell sounds. I will not need to run a heater in my studio this winter either.
I'm a citizen of the earth and a parent as well. I don't ignore my responsibilities. You should meet my children. They are two of the most well balanced people I know.
Do you have a design that sounds as good and one that those of us who are still learning can build? I'd love to see it. Maybe we can get Al to come up with another PCB GB.
Blessings, Terry
Guys, stick to the topic.
Audioguru, your comments, though understandable, were phrased in a very offensive manner, and your language terrible. If I find you making such comments in the future, in any thread, I will put you in the sinbin.
al/disappointed he had to put his mod hat on in this thread.
Upupa Epops said:When you are talking about PCB, look at my " configuration "... isn't it optimal ? It is my different amp, but you can take inspiration... All is " in one " and you can't make shorter traces ...
THAT is big brother...
Pavel, Pleeaaase let us know when it is finished. I would love to build one. Yes I know you stated it would not be for beginners, but as long as it has a complete BOM I can construct it and make you proud.
I've been thinking (oh no, not that) and it seems to me that another approach to reducing the power consupmtion of the output stage rather than switching the bias, it to gate the power to the transitor pairs and/or drivers with relays. This way, you could elect to run one, two or three pairs of output transitors based on say a front panel switch, but for lower power use, you would still have plenty of usuable output current, and maintain a high bias state, which some seem to think sounds subjectively better.
Comments, suggestions?
Comments, suggestions?
ok
What would you disconnect, the collector (from VDC) and base (from Vin) of the output transistors? Would you need to disconnect the output terminal from the emitters...hmm... probably not.
I ran this through K-amps spreadsheet.
Lets suppose I did this to my amp, in my situation, going from 4 output pairs to 2 pairs per channel would reduce power consumption from 146 to 73 WPC and power output from 66WPC to 16 WPC class A, class B power is still up at 66WPC. While overall power consumption is lower, efficiency drops from 45% to 22%.
So to me it looks like it would work. Good idea.
But I'm not an expert, so lets see what the smart people have to say....
What would you disconnect, the collector (from VDC) and base (from Vin) of the output transistors? Would you need to disconnect the output terminal from the emitters...hmm... probably not.
I ran this through K-amps spreadsheet.
Lets suppose I did this to my amp, in my situation, going from 4 output pairs to 2 pairs per channel would reduce power consumption from 146 to 73 WPC and power output from 66WPC to 16 WPC class A, class B power is still up at 66WPC. While overall power consumption is lower, efficiency drops from 45% to 22%.
So to me it looks like it would work. Good idea.
But I'm not an expert, so lets see what the smart people have to say....
Im not sure what I am missing, but for some reason I thought the there was a linear relationship between output pais and power output, e.g., ~17 Watts per pair. Thanks for the input guys, switches are certainly cheaper and easier to implement.
It occured to me, that with my KSA-50s, I almost never use more than the second bias level when listening (~10W), so a single pair of output transistors would probably suffice most of the time and not even require the fan to be on. In the simplest implementation, three switches could be employed with the second one tied to the fan circuit, assuring propper dissipation at moderate and higher power use. Does anyone have a line on a source for the Hafler sinks w/ fan that Terry is using?
Thanks again for the hard work on this project guys. I look forward to this coming together sometime this winter.
It occured to me, that with my KSA-50s, I almost never use more than the second bias level when listening (~10W), so a single pair of output transistors would probably suffice most of the time and not even require the fan to be on. In the simplest implementation, three switches could be employed with the second one tied to the fan circuit, assuring propper dissipation at moderate and higher power use. Does anyone have a line on a source for the Hafler sinks w/ fan that Terry is using?
Thanks again for the hard work on this project guys. I look forward to this coming together sometime this winter.
switching off output devices opposed to switching bias level per device will gain you nothing.
Instead, current capability is reduced as power supply voltage remains the same.
Damping factor is reduced, and voltage loss across the emitter resistors increases at the same output current because the same current will go through fewer emitter R's.
On top of that you'll have a relay contact in the powerlines, which requires a high current relay. High current relays with decent longevity and low contact resistance are pricy.
Or, a relay contact in the signal route, between output of the driver and base of the output devices.
In the long run you'll risk a loss of sound quality, you'll need the best relays there too.
By switching output devices one pair will have a higher risk of failure than the other, as average temperature between two pairs will differ.
The cost is higher than a single small relay for switching different value Vbe multiplier resistors, and there is no gain, only loss.
The relationship between bias and power output in class A is quadratic, not linear.
Switching different voltage output stage transformers is a better choice, keeping the front end on the nominal voltage.
Reduce voltage by half and dissipation goes down 50 %, efficiency remains the same.
Some class A amplifier designs used this trick to match different impedance loads.
Instead, current capability is reduced as power supply voltage remains the same.
Damping factor is reduced, and voltage loss across the emitter resistors increases at the same output current because the same current will go through fewer emitter R's.
On top of that you'll have a relay contact in the powerlines, which requires a high current relay. High current relays with decent longevity and low contact resistance are pricy.
Or, a relay contact in the signal route, between output of the driver and base of the output devices.
In the long run you'll risk a loss of sound quality, you'll need the best relays there too.
By switching output devices one pair will have a higher risk of failure than the other, as average temperature between two pairs will differ.
The cost is higher than a single small relay for switching different value Vbe multiplier resistors, and there is no gain, only loss.
The relationship between bias and power output in class A is quadratic, not linear.
Switching different voltage output stage transformers is a better choice, keeping the front end on the nominal voltage.
Reduce voltage by half and dissipation goes down 50 %, efficiency remains the same.
Some class A amplifier designs used this trick to match different impedance loads.